DESCRIPTION (provided by applicant: The proposed work develops further the statistical energy landscape approach to protein folding dynamics and structure prediction. The specific aims are: 1) to elucidate molecular origins of protein energy landscape topography, emphasizing the role of the solvent and side chain degrees of freedom; 2) to elucidate the microscopic origins of kinetic barriers to folding and of protein metastability; and 3) to further develop rigorously based statistical mechanical algorithm for structure prediction that account for solvent and side chain degrees of freedom. Energy landscape theory provides mathematical techniques for characterizing, in probabilistic terms, the energies of the ensembles of partially folded protein configurations and the dynamics of interconverting between them. Both analytical and computer simulation approaches are proposed that will provide quantitative estimates for the role individual amino acids play, both in guiding the protein to its native state, and in impeding that flow. Changes in folding kinetics affect protein trafficking and are thought to be involved in the pathogenisis of many diseases including Alzheimer's disease, Type II diabetes and cystic fibrosis. The improved ability to predict accurate three dimensional protein structures from sequence will be of great value in generally making use of data obtained from both human and bacterial genome sequencing projects.